field initializer should win over implicit optional-argument null

[floitschG]

given the following class A, an instantiation of the form new A() should have its field initialized with 499.
===
class A {
  var x = 499;
  A([this.x]);
}
===
Currently the x-field is initialized with null, since [this.x] must be interpreted as if it was [this.x = null].

[gbracha]

"given the following class A, an instantiation of the form new A() should have its field initialized with 499."

I assume this is the behavior you are asking for, as opposed to the specified behavior. This is very much a matter of opinion, and so I'm changing this to a feature request.

I would rather issue a warning in such cases (or even make it an error).

I was very much against allowing initializing formals as optional parameters in the first case. Each time you try and push this little sugar further, you get into a deeper mess than before. I am against this feature request.

The this.x syntax started as a simple sugar for a common case. Every time someone has a more complex case, the answer should be that they should revert to the unsugared form - not that the sugar should have more involved rules.

---

Removed Type-Defect label.
Added Type-Enhancement, Accepted labels.

[gbracha]

Issue #2159 has been merged into this issue.

[gbracha]

Set owner to @gbracha.

[whesse]

Just to clarify, would the unsugared case be

  A([this.x = 499]);
or
  A([x_param]) : x = (x_param == null)? 499 : x_param;
?

Also, assume x is final, since that is the tricky case.

[floitschG]

I would expect A([this.x = 499]).
When entering the function without a value you basically ask if the parameter (which in this case is a field) has a default value.

[gbracha]

class A {
  var x = 499;
  A([x]){ this.x = x};
}

Or

class A {
  final x;
  A([xp = 499]): x = xp;
}

---

Added WontFix label.